U.S. patent number 8,370,109 [Application Number 12/454,519] was granted by the patent office on 2013-02-05 for machine vibration baseline synthesizer.
This patent grant is currently assigned to DLI Engineering Corporation. The grantee listed for this patent is Laurent H. LaPorte, David J. Van Dyke, William A. Watts. Invention is credited to Laurent H. LaPorte, David J. Van Dyke, William A. Watts.
United States Patent |
8,370,109 |
Van Dyke , et al. |
February 5, 2013 |
Machine vibration baseline synthesizer
Abstract
Machine vibration baselines are synthesized without averaging
historical vibration data. Instead machinery configuration
information is used, including data about the machine train, the
machine's rotating components, and the machine speeds. For each
component being considered, a plurality of parameters is identified
for deriving baseline portions. The parameters for one component
are used to identify baseline vibration amplitude floor values,
vibration amplitudes at frequencies corresponding to various
harmonics of a rotational speed, and vibration amplitudes
corresponding to forcing frequencies and their harmonics/sidebands.
Parameters for other components also may be used to identify
amplitudes at harmonic frequencies and forcing frequencies.
Inventors: |
Van Dyke; David J. (Bainbridge
Island, WA), Watts; William A. (Bainbridge Island, WA),
LaPorte; Laurent H. (Lynnwood, WA) |
Applicant: |
Name |
City |
State |
Country |
Type |
Van Dyke; David J.
Watts; William A.
LaPorte; Laurent H. |
Bainbridge Island
Bainbridge Island
Lynnwood |
WA
WA
WA |
US
US
US |
|
|
Assignee: |
DLI Engineering Corporation
(Poulsbo, WA)
|
Family
ID: |
41342721 |
Appl.
No.: |
12/454,519 |
Filed: |
May 18, 2009 |
Prior Publication Data
|
|
|
|
Document
Identifier |
Publication Date |
|
US 20090292505 A1 |
Nov 26, 2009 |
|
Related U.S. Patent Documents
|
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
Issue Date |
|
|
61128125 |
May 20, 2008 |
|
|
|
|
Current U.S.
Class: |
702/184; 702/33;
702/34; 702/35; 702/189; 702/182; 702/183; 702/66; 702/56; 702/54;
702/185 |
Current CPC
Class: |
G01M
13/028 (20130101); G01H 1/00 (20130101) |
Current International
Class: |
G06F
11/30 (20060101) |
Field of
Search: |
;702/184 |
References Cited
[Referenced By]
U.S. Patent Documents
Other References
Computational Systems, Inc. Reference Manual, pp. 2-24 to 2-31;
Released to US customers as early as May 19, 2007. cited by
applicant.
|
Primary Examiner: Teixeira Moffat; Jonathan C
Assistant Examiner: Park; Hyun
Attorney, Agent or Firm: Koda; Steven P.
Claims
What is claimed is:
1. A method of synthesizing a vibration baseline for a machine for
use in diagnosing vibration of the machine without using historical
vibration data measurements of said machine nor an identical
machine, the method being performed by a computing system having a
processor and memory, comprising: receiving into memory machine
configuration data; processing by the processor the machine
configuration data to identify a prime component, one or more other
components, and one or more forcing frequency indicators of the
machine for which the vibration baseline is to be synthesized;
accessing a baseline specification database stored in memory to
obtain parameters for deriving the vibration baseline from the
identified prime component, one or more other components, and one
or more forcing frequency indicators, wherein the baseline
specification database comprises a plurality of parameters for each
of a plurality of machine component categories, the plurality of
parameters for at least one category comprising multiple harmonic
frequencies and multiple forcing frequencies, wherein all content
of the baseline specification database is derived without any of
said historical vibration data measurements; formulating said
vibration baseline for the machine without using historical
vibration data measurements of said machine, said formulating
comprising; setting by the processor first vibration amplitude
values of the machine vibration baseline at prime component
harmonic frequencies, wherein the prime component harmonic
frequencies are determined at least in part from content of the
baseline specification database according to the identification of
the prime component; setting by the processor second vibration
amplitude values of the machine vibration baseline at other
component harmonic frequencies based at least in part from content
of the baseline specification database according to the
identification of said one or more other components; and setting by
the processor a third vibration amplitude value of the machine
vibration baseline at a forcing frequency based at least in part
from content of the baseline specification database according to
one of said one or more forcing frequency indicators.
2. The method of claim 1, further comprising: setting by the
processor respective vibration amplitude floor values for at least
a first and a second frequency range of the baseline based at least
in part from content of the baseline specification database
according to the identification of the prime component.
3. The method of claim 1, wherein said step of receiving comprises
receiving into memory machine train data, component data and
machine speed data, and wherein said setting first vibration
amplitude values are based at least in part from content of the
baseline specification database according to the identification of
the prime component and the received machine speed data.
4. The method of claim 1, further comprising receiving into memory
a test location, wherein the prime component is identified based at
least in part on the test location.
5. The method of claim 1, wherein said one or more forcing
frequency indicators comprises blade pass.
6. The method of claim 1, wherein said one or more indicators of
vibration comprises gear mesh.
7. The method of claim 1, wherein said one or more indicators of
vibration comprises a bearing tone.
8. The method of claim 1, wherein said one or more indicators of
vibration comprises a belt rate.
9. The method of claim 1, wherein said one or more indicators of
vibration comprises flow noise.
10. The method of claim 1, wherein said one or more indicators of
vibration comprises a harmonics decay pattern.
11. The method of claim 1, wherein said one or more indicators of
vibration comprises a sideband decay pattern.
12. The method of claim 1, further comprising accessing with the
processor, based upon the identification of the prime component, a
plurality of parameters to be used for deriving portions of the
baseline, wherein said plurality of parameters are used for said
setting said first vibration amplitude values and said third
vibration amplitude value.
13. The method of claim 1, further comprising shaping with the
processor the baseline in the vicinity of said first vibration
amplitude values and said second vibration amplitude values.
14. The method of claim 1, further comprising processing with the
processor the machine configuration data to identify a first code
for the prime component, respective second codes for each one of
the one or more other components, and respective third codes for
each one of the identified one or more forcing frequency
indicators, wherein the first code, respective second codes, and
respective third codes are used to access vibration parameters from
the baseline specification database for use in synthesizing the
machine vibration baseline.
15. A method for screening a machine with a diagnostic system
having a processor, memory, and one or more sensors, comprising:
identifying by the processor machine train, components, machine
speed and test location for the machine based on input data;
configuring with the processor a diagnostic system for screening
the machine; measuring with said one or more sensors actual machine
vibration data generated by the machine; generating with the
processor the machine vibration baseline from said historical
vibration data when available; synthesizing by the processor the
machine vibration baseline without said historical vibration data
when said historical vibration data is not available; and analyzing
with the processor the actual machine vibration data, wherein said
analyzing comprises accessing the machine vibration baseline; and
wherein said synthesizing synthesizes the machine vibration
baseline within a frequency range for the test location without
averaging historical vibration data from either one or both of the
machine and an identical machine, and comprises: accessing a
baseline specification database stored in memory to obtain
parameters for deriving the machine an vibration baseline, wherein
the baseline specification database comprises a plurality of
parameters for each of a plurality of machine component categories,
the plurality of parameters for at least one category comprising
multiple harmonic frequencies and multiple forcing frequencies,
wherein all content of the baseline specification database is
derived without any of said historical vibration data; and setting
a baseline vibration amplitude value for at least each one of a
harmonic frequency and a forcing frequency that are based at least
in part on identification of a prime component.
16. The method of claim 15, in which said synthesizing comprises:
receiving into memory machine configuration data; processing with
the processor the machine configuration data to identify the prime
component, one or more other components, and one or more forcing
frequency indicators; setting by the processor a first vibration
amplitude value of the machine vibration baseline at said harmonic
frequency; setting by the processor second vibration amplitude
values of the machine vibration baseline at harmonic frequencies
within the baseline based at least in part on the identification of
said one or more other components; and setting by the processor a
third vibration amplitude value of the machine vibration baseline
at said forcing frequency based at least in part on one of said one
or more forcing frequency indicators associated with the prime
component.
17. The method of claim 16, further comprising: setting by the
processor respective vibration amplitude floor values for at least
a first and a second frequency range of the baseline based at least
in part on the identification of the prime component.
18. The method of claim 16, further comprising identifying with the
processor, based upon the identification of the prime component, a
plurality of parameters to be used for deriving portions of the
baseline, wherein said plurality of parameters are used for said
setting said first vibration amplitude values and said third
vibration amplitude value.
19. An apparatus for synthesizing a machine vibration baseline
within a frequency range for a test location of a machine without
averaging historical vibration data from either one or both of the
machine and an identical machine, comprising: means for inputting
machine configuration data; processing means for processing the
machine configuration data to identify a prime component, one or
more other components, and one or more forcing frequency indicators
of the machine for which the vibration baseline is to be
synthesized; a baseline specification database comprising a
plurality of parameters for each of a plurality of machine
component categories, the plurality of parameters for at least one
category comprising multiple harmonic frequencies and multiple
forcing frequencies, wherein all content of the baseline
specification database used for synthesizing a first machine
vibration baseline is not based on any historical vibration data
obtained from the machine nor an identical machine; processing
means for setting first vibration amplitude values of the first
machine vibration baseline at prime component harmonic frequencies,
wherein the prime component harmonic frequencies are determined at
least in part from content of the baseline specification database
according to the identification of the prime component; processing
means for setting second vibration amplitude values of the first
machine vibration baseline at other component harmonic frequencies
based at least in part from content of the baseline specification
database according to the identification of said one or more other
components; and processing means for setting a third vibration
amplitude value of the first machine vibration baseline at a
forcing frequency based at least in part from content of the
baseline specification database according to one of said one or
more forcing frequency indicators.
20. The apparatus of claim 19, further comprising: processing means
for setting respective vibration amplitude floor values for at
least a first and a second frequency range of the baseline based at
least in part on the identification of the prime component.
21. The apparatus of claim 19, further comprising memory for
storing vibration amplitude values for rotational speed harmonics,
forcing frequencies and other vibration parameters from which said
first vibration amplitude values, said second vibration amplitude
values, and third vibration amplitude value are derived, wherein
said values stored in memory are obtained without averaging
historical vibration data from said machine.
22. A diagnostic testing system for screening a machine,
comprising: means for receiving machine train, components, machine
speed and test location data for the machine; a baseline
specification database comprising a plurality of parameters for
each of a plurality of machine component categories, the plurality
of parameters for at least one category comprising multiple
harmonic frequencies and multiple forcing frequencies, wherein all
content of the baseline specification database used for
synthesizing a first machine vibration baseline is not based on any
historical vibration data obtained from the machine nor an
identical machine; sensors for measuring actual machine vibration
data generated by the machine; a machine vibration baseline
generator for generating a second machine vibration baseline from
said historical vibration data when said historical vibration data
is available for either the machine or an identical machine, and
for synthesizing said first machine vibration baseline without said
historical vibration data when said historical vibration data is
not available, said first machine vibration baseline being within a
frequency range for the test location and being based on content
from the baseline specification database as selected based upon the
received components data and without averaging historical vibration
data from either one or both of the machine and an identical
machine, wherein said synthesizing comprises setting a baseline
vibration amplitude value for at least each one of a harmonic
frequency and a forcing frequency that are based at least in part
on parameters read from the baseline specification database and
identification of a prime component; and processing means for
analyzing the actual machine vibration data, wherein said
processing means accesses the synthesized first machine vibration
baseline.
23. The system of claim 22, wherein said machine vibration baseline
generator comprises: means for inputting machine configuration
data; processing means for processing the machine configuration
data to identify a prime component, one or more other components,
and one or more forcing frequency indicators; processing means for
setting first vibration amplitude values of the first machine
vibration baseline at prime component harmonic frequencies, wherein
the prime component harmonic frequencies are determined at least in
part from content of the baseline specification database according
to the identification of the prime component; processing means for
setting second vibration amplitude values of the first machine
vibration baseline at other component harmonic frequencies based at
least in part from content of the baseline specification database
according to the identification of said one or more other
components; and processing means for setting a third vibration
amplitude value of the first machine vibration baseline at a
forcing frequency based at least in part from content of the
baseline specification database according to one of said one or
more forcing frequency indicators.
24. The apparatus of claim 23, further comprising: processing means
for setting respective vibration amplitude floor values for at
least a first and a second frequency range of the baseline based at
least in part on the identification of the prime component.
25. The apparatus of claim 23, further comprising memory for
storing vibration amplitude values for rotational speed harmonics,
forcing frequencies and other vibration parameters from which said
first vibration amplitude values, said second vibration amplitude
values, and third vibration amplitude value are derived, wherein
said values stored in memory are obtained without averaging
historical vibration data from said machine.
Description
FIELD OF THE INVENTION
The present invention generally relates to methods and apparatus
for monitoring the condition and reliability of machinery, and more
particularly to methods and apparatus for synthesizing vibration
profiles of machinery, such as may be used for screening out or
troubleshooting faulty machines.
BACKGROUND OF THE INVENTION
Industrial machines may vary in their complexity, but typically
include several rotating elements, such as bearings, shafts and
gears. Where there is rotation there generally is some degree of
machine vibration. Vibration occurs due to imbalance, misalignment
of shafts, and bearing wear. Every machine has some level of
vibration that is characteristic of its operation. In particular,
properly maintained machines are designed to exhibit not more than
a reasonable degree of vibration under normal operating conditions.
Excessive levels of vibration indicate machine problems that may
lead to unsatisfactory operation and eventual breakdown.
Machine malfunction can sometimes be detected by a change in the
vibration pattern of a machine. In today's plants where machines
are complex and/or large in number, engineers often rely on
advanced methods of identifying abnormal levels and patterns of
vibration in order to determine the condition of a machine. For
example, based upon knowledge of the rotational speed of individual
machine elements, machine maintenance personnel can monitor the
machine's vibration level at certain characteristic frequencies to
acquire an indication of the overall condition of the machine. As
the mechanical integrity of a machine element begins to degrade,
the vibration level associated with that element changes from its
normal characteristic level. To the trained machine maintenance
personnel such change may indicate that corrective action will soon
be necessary. By implementing a machine monitoring program, the
machine's vibration levels can be measured on a regular schedule,
and early detection of abnormal machine operation is possible. With
such early warning, repair of the machine may be scheduled well
before a machine breakdown and the associated work stoppage occurs.
In this manner, machine "down-time" may be scheduled well in
advance so as to minimize the impact on manufacturing
operations.
One challenge in predicting potential failures is determining what
vibration characteristics may be considered normal and safe--i.e.,
what is the machine's normal characteristic level of vibration.
Further, what are the variations in vibrations that fall outside
the normal and may indicate machine underperformance, component
wear, and ultimate machine breakdown. Establishing what is
considered "normal" operating vibration signatures for machinery
has been historically accomplished by selecting candidate machinery
vibration tests from identical machines without mechanical faults,
normalizing for speed variations, and numerically averaging them
together to obtain an "average" vibration signature. Vibration
spectra typically are collected at a number of locations on a given
machine. Specific spectral features in the measured data may
include harmonic families or difference families, which are
associated with certain types of machinery faults.
When a new, identical machine is installed, it may be monitored by
comparing its vibration profile with the historical average
vibration profile previously obtained for that machine. This
methodology is feasible and useful when a large number of
historical sets of vibration tests for groups of machines are
available. However, it is not feasible when there is a lack of such
abundant test data from a variety of machines in various
conditions. Furthermore, building average baselines requires
sufficient knowledge and experience in machinery vibration
signature analysis. In particular, the historical averaging
operation is subjective. A user needs to determine whether an
acquired candidate set of data is `averagable`--meaning whether it
is good quality data from an apparently fault free machine. Many
machine users do not have the skills or experience to judge whether
test data is suitable for establishing a baseline.
Accordingly, there is need of an alternative method of obtaining a
baseline vibration profile for machines, such as when data for
similar machines without faults is not available. Further, there is
a need to obtain vibration profiles that effectively define normal
and that may be efficiently applied to desired machinery to screen
out machines that are faulty or in need of maintenance, and to
troubleshoot machines identified as being faulty or in need of
maintenance. These and other needs are addressed by the inventions
described herein.
SUMMARY OF THE INVENTION
The present invention provides a method and apparatus for
synthesizing machinery vibration baseline spectra based upon
machinery descriptive information and previously compiled baseline
specifications data. The machinery descriptive information may
include, for example, machine train data, component data and
machine speed data. Baselines for a machine may be derived for a
specific test location and specific frequency range without
averaging historical vibration data from either one or both of the
machine and an identical machine.
A prime component is identified based at least in part on the test
location. Baseline specifications data relating to the prime
component are used to determine vibration amplitude floor values
for portions of the baseline, along with other vibration amplitude
values corresponding to various frequencies. The various
frequencies may include the rpm associated with the component,
harmonics of the rpm, forcing frequencies, and associated forcing
frequency harmonic and/or sideband frequencies. Baseline
specifications data relating to other components are similarly used
to determine amplitude peaks at other frequencies, such as at the
respective rpm associated with the other components, harmonic
frequencies of the rpm, and in some instances forcing frequencies
and associated forcing frequency harmonic and/or sideband
frequencies.
The invention will be better understood by reference to the
following detailed description taken in conjunction with the
accompanying drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
The invention is further described in the detailed description that
follows, by reference to the noted drawings by way of non-limiting
illustrative embodiments of the invention, in which like reference
numerals represent similar parts throughout the drawings. As should
be understood, however, the invention is not limited to the precise
arrangements and instrumentalities shown. In the drawings:
FIG. 1 is a graph of machine vibration baselines synthesized in
accordance with an embodiment of the present invention;
FIG. 2 is a graph of actual machine vibration data and the
synthesized vibration baselines of FIG. 1;
FIG. 3 is a data and control flow diagram for an apparatus for
synthesizing machine vibration baselines in accordance with an
embodiment of the present invention;
FIG. 4 is a flow chart of a method of synthesizing machine
vibration baselines in accordance with an embodiment of the present
invention;
FIG. 5 is a chart of category codes and potential fault codes for
various machine component configurations;
FIG. 6 is a data and control flow diagram for a system for
screening and diagnosing faults in accordance with an embodiment of
the present invention; and
FIG. 7 is a flow chart of a method for screening and diagnosing
faults, in accordance with an embodiment of the present
invention.
DETAILED DESCRIPTION OF ILLUSTRATIVE EMBODIMENTS
In the following description, for purposes of explanation and not
limitation, specific details are set forth, such as particular
devices, computers, networks, components, techniques, protocols,
software products, operating systems, development interfaces,
hardware, etc. in order to provide a thorough understanding of the
present invention.
However, it will be apparent to one skilled in the art that the
present invention may be practiced in other embodiments that depart
from these specific details. Detailed descriptions of well-known
devices, computers, networks, components, techniques, protocols,
software products, operating systems, development interfaces, and
hardware are omitted so as not to obscure the description of the
present invention.
Vibration analysis typically involves comparing measured vibration
from a machine with a standard benchmark, datum, or baseline. Some
standards are simple, like single values for broad ranges of
frequencies. Other standards may consider the machine type and
assign values to certain frequency regions based on the machine
speed. Full baseline vibration spectra can be quite complex with
values derived for a large number of discrete frequencies. For
example, there may be arrays of values spanning set frequency
ranges. A full test for a machine may include the collection of
12-18 spectra, such as when two frequency ranges are used in 3
directions at 2 or 3 test locations on a machine.
Vibration phenomena seen in spectra such as decaying harmonic
series, sidebands, or flow noise are known to vibration engineers.
But these nuances are hard to translate on per component and per
machine basis. Industry and machinery experience is often required
to develop alarm thresholds or baselines for automated systems. The
machine vibration baseline synthesizer of the present invention has
been developed to synthesize spectra vibration signatures of
various machine trains. A benefit of the synthesizer is that it
provides an "out of the box" reference that may be used in
automated diagnostic systems to screen out faulty machines and
accurately diagnose component specific problems. This is especially
beneficial for new machines, and new machine models, where
historical data is unavailable. Accordingly, rather than require
experienced personnel to determine a baseline average vibration or
threshold levels for alarms, an operator with little experience may
run the baselines synthesizer to screen for faulty machines and to
diagnose faults.
According to an embodiment of the present invention, a baseline
synthesizer processes machinery technical descriptions (e.g.,
machine train, components, machine speeds, and test locations) to
intelligently compute a set of baseline vibration spectra,
mimicking what a vibration test on an ideal fault-free machine of
that type would produce. The baseline synthesizer models the effect
of vibration generated from background, internal machinery
components and vibration excited by nearby connected components. It
generates signatures for bearing locations in multiple axes and
multiple frequency ranges.
For example, to assist in the determination of machine problems, it
is very helpful to have vibration data from each measurement point
along three axes--axial, radial, and tangential. Axial is the
direction parallel to the shaft in question, radial is the
direction from the transducer to the center of the shaft, and
tangential is 90 degrees from radial, tangent to the shaft.
FIG. 1 shows a graph 100 of baseline sets of vibrations derived in
accordance with an embodiment of the present invention for a given
machine. In this example, the baselines are for a motor driving a
fan via a belt having a belt ratio of 1.563:1 at a motor speed of
1780 revolutions per minute. In the given example, baselines may be
derived for each of multiple locations on the machine. For the
motor of this example, it may be desirable to synthesize baselines
for vibrations at the motor free end, motor drive end, the fan
drive end, and the fan free end. FIG. 1 shows the axial, radial and
tangential baselines 102, 104, 106 for vibrations in a low
frequency range at the motor free end. The horizontal axis
corresponds to frequency. The vertical axis corresponds to
vibration amplitude. Baselines may be synthesized for other
frequency ranges and locations in accordance with a desired
implementation. Further, baselines may be derived for more complex
machines having many rotating parts.
FIG. 2 shows a graph 120 of another set of baselines 103, 105, 107
overlaid onto corresponding sets 112, 114, 116 of actual vibration
data. The horizontal axis corresponds to frequency. The vertical
axis corresponds to vibration amplitude. The actual vibration data
was sensed on an actual machine at a given test location and for a
specified frequency range. The synthesized baselines 103, 105, 107
are derived for the same test location and the same or an
overlapping frequency range. Taking the baselines as being maximums
for normal vibration, any peaks where the actual vibration data
exceeds the baseline may indicate abnormal vibration. For example
the actual axial vibration data 112 at the motor free end in the
illustrated example has 5 peaks 118 exceeding the baseline 103.
Also, there are several peaks 122 exceeding the baseline 105 for
the radial vibration data. With regard to the tangential vibration
data in this example, all the actual vibration data peaks within
the frequency range of the synthesized baseline 107 are below the
baselines 107. In other examples, the positions of the synthesized
baseline vibration peaks at the baselines 103, 105, 107 may vary.
Of course, the actual vibration data also may vary at different
times and for different specimens of the identical machine. In
various implementations, the synthesized baselines 103, 105, 107
may serve as alarm limit envelopes, or as a starting point for
analysis of fault screening and fault diagnosis.
In some instances, baselines may be derived using machine
descriptions which are less than complete. Although the baselines
can still provide valuable screening information, in such instances
there may be false positives. In the current example, the number of
fan blades may be unknown. In such instance, there may be peaks in
the actual vibration data exceeding the baseline, because the
actual number of fan blades was not considered. A technician
viewing the result may quickly recognize the pattern of the actual
vibration data as corresponding to a fan with `y` number of blades
and recognize that one or more peaks are due to the unaccounted
blade(s), rather than being due to a faulty machine. Accordingly,
the results may be considered normal.
Vibration Baseline Synthesis Data and Control Flow
A baseline synthesizer produces a nominal baseline by applying
standards developed through empirical and analytical methodology.
The purpose is to reliably and effectively estimate baseline
spectra for various machine types. Such baselines may be derived
without the need for historical vibration data obtained from the
same machine or identical machines of the same type. In some
applications, the baselines may be derived on the fly during
machine screening and diagnostic testing. In other implementations
the baselines may be derived and stored for later use.
FIG. 3 shows a data and control flow 150 for a baseline synthesizer
152 according to an example embodiment of the present invention.
The baseline synthesizer 152 may be implemented in software and
executed on a computing system, such as a personal computer, laptop
computer, handheld computer, workstation, server, embedded
computing device, or other computing device. In some embodiments
the baseline synthesizer 152 may be implemented in a network of
computing systems. Various memory devices, processors, displays,
input devices, and output devices may form all or portions of a
computing system implementing the baseline synthesizer 152.
According to an example embodiment of the present invention, the
machine vibration baseline synthesizer 152 computes a unique set of
spectra for each associated component and bearing within a machine
train, and combines the data into a baseline for a given test
location and frequency range. The synthesizer 152 receives inputs
which identify the machine configuration, (e.g., machine train,
components, machine speeds), along with one or more test location
and an output frequency range for which baselines are to be
derived. For example, an operator may go through an automated
interview process or another method to provide input 160 for
controlling the baseline synthesizer 152. Various data may be
included among or derived from the inputs 160, such as run
configuration data 153, component codes 154, and fault codes 156.
To process the inputs, a baseline specification database 158 may be
accessed.
An operator may specify run configuration data 153, such as the
machine model and/or type (e.g., and thus, the machine train
information), one or more locations that are to be used on the
machine as test locations, and one or more frequency ranges for
which the baselines are to be derived. Baselines may be derived for
each test location identified by the operator for each frequency
range. For example, machine vibration analysis is often performed
for both a low frequency range (e.g., 10.times. the main shaft rpm)
and a high frequency range (e.g., 100.times. the main shaft rpm),
although the operator may specify a specific or other frequency
range.
The operator also may provide information about the machine
configuration, (e.g., motor type, couplings, bearing types,
component orientation (vertical/horizontal), component location).
For example, by going through an automated question-answer session
with the operator/user, a set of component codes 154 may be
derived, which identify the respective components to be considered
in the synthesis process. The component codes 154 implicitly define
many aspects of the machinery being tested. For example, a
component code may correspond to a motor with ball bearings driving
a fan using a belt. The same physical motor driving a fan with no
belt, however, may correspond to a different component code. The
specific component code 154 may be obtained by looking up the
component code in a database based upon the user responses, or by
building a component code based on specific input responses.
Component codes 154 may be input to the baseline synthesizer 152
for each component for which vibration contributions are to be
factored. In other embodiments, another method may be used for
identifying a component to be considered in the baseline synthesis
process.
Below is a sample listing of components for which component codes
154 may be included. One of ordinary skill will appreciate that
varying numbers and types of components may be included, in
accordance with a given implementation of the baseline synthesizer
152.
Component Identification: AC Line Frequency (50/60 Hz) Motors
Horsepower AC (VFD) DC Motor Cooling Fan Blades Flexible Couplings
Close Coupled Units Gear Boxes: (Single/Double/Multi Speed) Gear
Tooth Counts Gear Ratio's . . . Belt Drives Belt Rate Sheave
Diameters and Distances . . . Pumps Centrifugal Piston . . . Fans
Axial Centrifugal . . . Compressors Piston Rotary Screw . . .
Bearings Rolling Element/Journal Bearing Locations (tested
locations) . . .
Also associated with the machine configuration information provided
by operator may be fault codes 156 or other identifying codes. The
fault codes 156 are codes for identifying or otherwise indicating
forcing frequencies to be accounted for in the baseline synthesis
process. In an example embodiment, a fault code may be derived from
the machine configuration inputs provided by the operator. By
providing information about a component, the machine running speed,
and a relation between the component and a machine running speed
(e.g., gear ratio), information pertaining to vibration phenomena
may be identified. For example, a specific fault code and/or other
forcing frequency parameter information may be identified
automatically. In other embodiments, the forcing frequencies or
other vibration phenomena may be indicated directly or by a method
other than by an identification code.
Consider an example in which a fault code relevant to component
category T (which may for example be for a positive displacement
pump), may be "PP" (i.e., pump pistons), and identify that there
are 9 elements, (PP=9). If the fault codes for the machine type
include a PP code and the prime component is a Category T, then
that order (PP=9) is used as one forcing frequency FF to be
included in synthesizing the baseline. Several relevant forcing
frequencies may be possible for any one given component category.
Further, the fault codes may contain definitions for any number of
these forcing frequencies.
Various vibration phenomena creating known patterns in measured
spectra may be incorporated into the synthesis process. For
example, indicators of vibration phenomena relating to blade pass,
belt rate, gear mesh, flow noise, and adjacent components may be
incorporated, such as by using forcing frequency parameters. For
example, as the data for each component is processed, fault codes
determined according to the category of the component may be
accessed to identify the forcing frequency(ies) of the component
that are applicable for a given machine type. These forcing
frequencies may be stored as parameters in the baseline
specifications database 158. For example, the fault code and the
fault code value may determine a parameter and parameter value to
be included in synthesizing a baseline. The parameter values in
effect model the common source of vibration. These values are
obtained using empirical and analytical methodologies. Accordingly,
specific decaying patterns, noise floor levels, and/or sidebands
may be factored in based on the component and various forcing
frequencies.
Several forcing frequencies may be stored in the baselines
specification database 158 for a given category. Forcing
frequencies as used herein are the specific frequencies at which a
mechanical component will generate forces on a machine. Every
rotating component of any machine will produce forces in the
machine that will cause vibration at specific frequencies. The
specific forcing frequencies are related to the physical makeup of
a component. Each component may have one or more forcing
frequencies. One of the most significant forcing frequencies is
shaft rate, (i.e. revolutions per minute (RPM) of a shaft). This
forcing frequency arises due to the residual imbalance that a rotor
will always have. Such imbalance imparts a radial centripetal force
on the bearings, causing the structure to vibrate at the 1.times.
frequency, also known as the fundamental frequency. This is one of
the forcing frequencies. Other forcing frequencies are bearing
tones, which are characteristic of each bearing's geometry. Bearing
tones are due to forces generated by defects in the races and
rolling elements of the bearing itself. Another type of forcing
frequency includes the gear tooth-mesh frequencies that arise due
to the individual impacts of gear teeth against each other. The
tooth-mesh frequency is equal to the number of teeth on the gear
times the gear RPM. Another type of forcing frequency includes the
vane pass or blade pass frequencies. These are similar to tooth
mesh frequencies, and are equal to the number of vanes in an
impeller or the number of blades in a fan times the RPM. Each
forcing frequency will create a peak in the vibration spectrum. The
amplitude of the peak is dependent on the severity of the condition
that causes it. Thus the frequency where the peak occurs indicates
the type of problem and the amplitude of the peak indicates the
problem severity.
The baseline specification database 158 may be embodied in a file
structure, or other data structure. In an example embodiment, the
baseline specification database may include a plurality of
parameters for each of multiple categories. Each row in the
database (or line in the file), for example, may correspond to a
category. Each column may correspond to a parameter. Each component
may be of a certain category. Example categories may include one or
more categories of motors, gear boxes, bearings, compressors, fans,
pumps, belt drives, motor cooling fan blades, flexible couplings,
and close coupled units. Other categories may be defined. Thus,
inputs 160 that relate to a specific component may be used to
identify a component code, which in turn may lead to identification
of a category.
In various embodiments, the number of parameters and the specific
parameters may vary. In one embodiment there may be a common set of
parameters for each category. In another embodiment, there may be
differing sets of parameters associated with differing categories.
In a specific embodiment of the baseline specifications database
158, there are several parameters for each given category, although
not all parameters need be used, or include valid data. Following
is a list of an example list of parameters stored for each category
in an example embodiment: 1X, 2X, 3X, 4X, 5X, 6X, 7X, 8X, 9X, 1xFF,
2-3xFF, 4-12xFF, fixed freq, floor1, floor2, floor3, floor4. With
regard to parameters, X refers to the main shaft RPM for the
categorized component. Thus, 1X to 9X refer to harmonics of the
main shaft. "FF" stands for fault code forcing frequency and may
refer to any of various forcing frequencies, such as for motor
bars, pump vanes, compressor pistons, gear mesh, etc. The specific
fault code determines the type of forcing frequency that is stored
in the baseline specifications database for a given category.
The floor values may be defined as follows: floor1 covers 0-2X
rotational rate, floor2 covers 2X-10X, floor3 covers 10X-40X, and
floor4 covers 40X-100X, where X refers to the shaft rate of the
categorized component. In some embodiments, additional parameters
may be included, such as for forcing frequencies of sidebands.
Specifically, an amplitude for a fundamental forcing frequency that
generates sidebands may be specified as one parameter. In addition,
parameters may be included for vibration amplitudes at various
sidebands, (e.g., first sidebands, second sidebands, twelfth
sidebands). The spacing of the sidebands may be derived from the
machinery descriptive information provided by the operator. For
example, the gear ratio for a component may determine at least in
part the sideband spacing for sideband parameters stored for a
given component category.
The value stored in the baseline specifications database 158 for a
given parameter may be a vibration amplitude for a specific order
of frequency. Specifically, the frequency for a given parameter may
be based upon the rotation rate, X, and the various orders of X, or
upon a forcing frequency and various orders of the forcing
frequency, while the stored value corresponds to a peak amplitude
at that frequency. These amplitudes are precompiled and stored in
the baseline specifications database 158 using empirical and
analytical methodologies.
Given the component codes 154 which represent a model describing
the kind of machine, the fault codes 156 which define the known
vibration sources inside the machine, and the baseline
specification database 158 of parameters, the baseline synthesizer
152 constructs a useful estimate of the machine's vibration
signature. In particular, vibration amplitude floor values,
vibration amplitude peaks at operating speed harmonics, and
vibration amplitude peaks at forcing frequencies may be selectively
read or computed from entries in the baseline specifications
database 158. This estimate of the machine's vibration signature
resembles the average baseline derived from actual historical
vibration data, but is derived without the need for actual
historical vibration data. FIG. 1, as previously discussed, shows
an example of synthesized baselines 102, 104, 106 in graph format.
Such baselines were derived based upon the following input derived
from user input 160: test locations; machinery speed; arrangement
of machinery component types such as electric motors, centrifugal
pumps, couplings, fans, belt drives, gear trains, compressors, et
cetera; transmission type such as belts drive, gear drive or both;
and transmission ratio.
One of ordinary skill will appreciate that the synthesized
baselines 166 (see FIG. 3) may be output in various formats, and
that the specific output will vary depending on the machine for
which the baselines are being derived.
Vibration Baseline Synthesis Method
FIG. 4 shows processes performed as part of a method 200 for
synthesizing machine vibration baselines, according to an example
embodiment of the present invention. The method 200 may be
performed for any given machine, for various locations on the
machine, for various frequency ranges, and encompass the impacts
from various components. At step 202, machinery descriptive
information (e.g., drive train, components, machine speeds) and run
configuration data (e.g., test location, frequency range), may be
obtained from an operator (or another source, such as a diagnostic
screening system). At step 204, the machinery descriptive
information may be processed to derive component data and forcing
frequency data. For example, the component codes 154 and the fault
codes 156 may be derived, such as by a process described above with
regard to the data and control flow of FIG. 3. The component codes
and fault codes form a simple model of the machine for which
vibration baselines are being derived. Each test location
corresponds to a location on the machine for which baselines are to
be synthesized. A given test location corresponds to the location
where a sensor would be placed if measuring actual vibration data.
The frequency range determines the output range for the synthesized
baselines, (e.g., baselines for a low frequency range; baselines
for a high frequency range; baselines for a specific frequency
range). By including a test location and frequency range, the
synthesized baselines may be used for comparing and analyzing
actual machine vibration data over the specified frequency range
obtained from a sensor at the test location. Generally, the
specified location is at a bearing, such as a driven end or free
end of a shaft. However, other locations may be specified.
At step 206, the test location, drive train and components data are
examined to identify the component that is to be treated as a prime
component for the analysis at the test location. For example, the
component at or near the location that is expected to impart the
most vibration may be selected as the prime component.
At step 208, parameters may be loaded from the baselines
specifications database 158 for the prime component. For example, a
category of the prime component may be identified, and the data for
such category retrieved from the baselines specifications database
158. In some embodiment, the category may be looked up in a
database based upon the specific component code. In other
embodiments the components code may logically relate to the
category and be parsed or otherwise processed to identify the
category. FIG. 5 shows an example table relating components 172 to
the category codes 174, along with potential fault codes. All the
data used to determine the category code is obtained among the
inputs 160 from the machine operator. One of ordinary skill will
appreciate that the data listed in FIG. 5 is not an exhaustive list
of components, categories, or fault codes. The specific components,
codes and number of codes may vary.
In a specific embodiment, a baseline may be formed as a set of
array of elements. For example, a 500 element array may be created
for a given baseline. At step 210, an array of elements
corresponding to axial, radial and tangential frequencies may be
constructed for a select frequency range. A starting point may be
to set vibration floor values for elements within specific
frequencies. A floor value is the minimum amplitude value in a
specific frequency region for considering vibration significant.
Amplitudes below the floor value of a specific region are not
considered significant (e.g., correspond to normal vibration
levels). Each floor value corresponds to a given frequency range.
Thus, over the specified frequency range for the axial baseline,
one floor value may correspond to a first portion of the baseline,
a second floor value may correspond to a second portion of the
baseline, and so on for the entire frequency range of the baseline.
The floor values are obtained from the parameters read from the
baselines specifications file 156 for the category of the prime
component. Specific noise floor values at 0-2X, 2x-10x, 10x-40x and
40x-100x may be used, where x corresponds to the running speed of
the prime component. Baselines may be started for each of the
axial, radial and tangential baselines in the same manner using the
floor values.
At step 212, vibration amplitudes may be set for specific harmonic
frequencies relating to the prime component. For example,
amplitudes may be stored in the baselines specification database
158 for the prime component's category for various orders of
harmonics. An amplitude may be read for the 1X parameter and be set
at the array element corresponding to the specific running speed
frequency (i.e., 1X) of the prime component. Note that although the
running speed may vary at different times, the amplitudes are
associated with frequency orders of the running speed. Similarly,
for each of the harmonics of the running speed, other amplitudes
may be read from the baseline specification database 158 and set as
the vibration amplitudes for the corresponding array elements.
At step 214, vibration amplitudes may be set for various forcing
frequencies associated with the prime component. For example,
forcing frequencies, or fault codes, may have been identified at
step 204. Consider an example, in which a specific fault code `PP`
(i.e., pump pistons) is identified, along with a value for the
fault code, `9` for the prime component. The fault code `PP`
indicates that the parameters for the forcing frequencies are to be
accessed for the prime component category in the baseline
specifications database. The fault code value, 9, indicates that
this value coincides with (and will replace) the 9.sup.th harmonic.
Thus, the value set in step 212 for the 9X frequency is to be
replaced with the value read for the 1XPP forcing frequency.
Further, the value set in step 212, if applicable, for the 18X
frequency is to be replaced with the 2XPP forcing frequency.
Similarly, the amplitude each multiple of the gth order frequency
is replaced with the corresponding nth order forcing frequency. In
some embodiments, fault codes also may encompass sideband forcing
frequencies. In such example, the sideband spacing is derived at
step 204 based on the machinery description and component data,
(e.g., may be derived from the gear ratio pertinent to the
component). Thus, a sideband fault code identifies that amplitudes
are to be set at sideband forcing frequencies. The value of the
fault code determines the center frequency. For example, it may be
5 and thus correspond to the array element set in step 212 for the
5.sup.th harmonic. The amplitude for the sideband may be read from
the baseline specifications database 158 for the component category
and set at the frequency corresponding to the 5X harmonic. Consider
an example in which the spacing is determined to be 0.5. The
amplitude for the +/-1 sideband forcing frequency may be read from
the baseline specifications database 158 at set at the frequency
4.5X and 5.5X. The amplitude for the +/-2 sideband forcing
frequency may be read from the baseline specifications database 158
at set at the frequency 4.0X and 5.0X. Note that for some of these
steps, the corresponding array element may be previously set to a
floor value, which now is being replaced. A similar process is
repeated for each of multiple sideband forcing frequencies
encompassed by the fault code. Accordingly, various harmonic and
sideband decay patterns may be considered according to the fault
codes, such as the 2x-3x forcing frequency harmonics (and/or
sidebands) and the 4x-12x forcing frequency harmonics (and/or
sidebands).
At step 216, amplitudes may be set for additional array elements
based upon the vibration impacts attributable to the other
components specified in the inputs at step 202, 204. These
components are considered non-prime components. The processes
performed are similar to those described with regard to steps 208,
212 and 214. For example, a category for each one of the non-prime
components may be obtained in the same manner as for the prime
component, and corresponding parameters for that category may be
read from the baseline specifications database 158. Like in step
212, additional amplitudes may be inserted into the array at the
various fundamental frequencies and harmonics for each non-prime
component. Also, like in step 214, vibration amplitudes may be set
for various forcing frequencies associated with the non-prime
components. For example, there may be fault codes that were
indicated for any one or more of these non-prime components.
Accordingly, the parameters corresponding to these fault codes may
be accessed to define additional vibration amplitude values to be
inserted into the array.
Also of note is that the shaft rate of the non-prime components may
differ from that of the prime component. Accordingly, the peaks for
the non-prime components are not likely to coincide with the peaks
caused by the prime component. Thus, when inserting these
attenuated peaks into the array the amplitude first is compared to
the array element amplitude (e.g., a floor value when not
coinciding with a prime component peak). If the non-prime component
peak is less than the floor value, it is ignored. If it exceeds the
floor value, then it replaces the floor value in the array. Shaping
of these additional peaks is performed using a similar shaping
process as for the prime component peaks.
The processes performed in step 218 differ from those in steps 212
and 214 with regard to the amplitude being stored in the array. In
particular, attenuation may be accounted for high frequency and low
frequency transmission loss across components. The impact of these
components is attenuated because the contribution is not on the
subject prime component, but instead is contributed via a
transferral from another component. Accordingly, the amplitude
stored in the array for a non-prime component is attenuated by a
nominal value. In one embodiment, the value of the appropriate
parameter read from baseline specification database for the
non-prime component category may be multiplied by an attenuation
factor or otherwise be decreased by a nominal attenuation. In other
embodiments, a specific attenuation may be derived for a given
component based upon a force transferrance path between the
component and the prime component.
At step 218, a shaping process may be performed on the array
elements to shaping the baselines. In an example embodiment
amplitudes for array elements adjacent to the specific array
element set at steps 212, 214 may be shaped by any of various
processes. For example, an amplitude at a specific array element
may be shaped over a predefined number of elements with the
specific frequency corresponding to the center element. The peak
may be given various shapes, such as a flat top, or be pointed. In
one embodiment, the peak drops by one third (of the difference
between the peak and the floor value) at the adjacent element on
each side, and by 2/3 at the next adjacent element on each side, so
that the third adjacent element is back down to the floor value for
that region of the baseline.
At step 218, the array may be formatted for output for each of the
axial, radial and tangential baselines corresponding to the test
location of the specified machine configuration. The array data may
be used for various purposes and may be output in various formats.
FIG. 1 shows an example of a graph of the axial, radial and
tangential baselines for a given test location of a given machine
over a given frequency range.
Machine Fault Screening and Diagnostic System
The machine vibration baseline synthesizer 152 may be used with or
as part of a machine screening system and/or a machine fault
diagnostic system 302. The fault diagnostic system 302 may be
implemented in software and executed on a computing system, such as
a personal computer, laptop computer, handheld computer,
workstation, server, embedded computing device, or other computing
device. In some embodiments the fault diagnostic system 302 may be
implemented in a network of computing systems. Various memory
devices, processors, displays, sensors, input devices, and output
devices may form all or portions of a computing system implementing
the fault diagnostic system 302.
The machine fault diagnostic system 302 may be used as part of a
predictive maintenance program to perform screening or other fault
diagnostic processes on a machine. Predictive maintenance utilizes
many different disciplines. By far the most significant portion
which helps to identify faults is periodic vibration analysis. Of
all the non-destructive testing that can be done on a machine, the
vibration signature provides the most pertinent information about
its inner workings. Accordingly, when screening identifies machines
that may have faults, fault diagnostics may be performed, such as
to identify source(s) of wear, imbalance, misalignment, mechanical
looseness or electrically-induced vibration. In some instances
baselines may be available that are statistically averages based on
historical data measured by sensors on the machine or an identical
machine of the same type. In such case, the fault diagnostic system
302 may use baselines derived from historical data. In other
instances, such as for new machines or new models of machines,
there may not be applicable historical data. Further, in some shops
the machine operator may not have the expertise to create a
statistical average of historical data, (e.g., may not be able to
normalize the data or otherwise make data for a given measurement
session comparable with other data). In these instances, the
machine fault diagnostic system 302 may request or access the
synthesized baselines derived according to an embodiment of this
invention.
FIG. 6 shows a data and control flow diagram for such a system 302
including the baseline synthesizer 152, in accordance with an
embodiment of the present invention. Like items and components are
given the same part numbers and provide the same or similar content
and function as described above for the baseline synthesizer flow
of FIG. 3. When performing machine screening, a vibration sensor
311 may be placed at each of one or more test locations 304. Such
locations are input to the baseline synthesizer 152 based upon the
user inputs 160. In some embodiments, these locations may be
detected and thus, directly input to the baseline synthesizer
without requiring an operator to specify the location.
When setting up the machine test 306, various frequency ranges and
sampling rates 308, and other test parameters also may be
specified. During the test, various vibration readings 310 are
input to the automated diagnostic system 302. The automated
diagnostic system 302 compiles, transforms and processes the
frequency readings to generate machine vibration spectra. These
vibration spectra may be compared with the synthesized baselines to
determine whether faults are present. FIG. 2 shows an example in
which actual machine vibration spectra 112, 114, 116 for axial,
radial and tangential vibration are compared with synthesized
baselines 103, 105, 107 for the same test location and the same or
an overlapping frequency range. In some cases, an alarm condition
or fault may be indicated when the actual data exceeds the
baseline, or when the actual baseline exceeds the baseline by a
threshold amount. In some cases, the pattern of the actual data
relative to the baselines may be indicative of a certain type of
fault. For instance, an increase at a 1X frequency of a motor
bearing may indicate an increasing imbalance condition. If
harmonics of 1X begin to appear, this is an indication of damage,
such as bearing clearance increases, looseness, or cracking of the
structure. Accordingly, a strong 1X vibration may indicate that the
rotor should be balanced, while the appearance of harmonics of 1X
may indicate that the bearing and surrounding structure should be
inspected for damage. The analysis of the actual vibration data in
comparison with the synthesized baselines may yield results 312
which are output, displayed, and/or stored.
Note that conventional diagnostic systems may use historical data
to form statistical average baselines 314 for comparison with
actual test data derived from readings 310. The system of this
invention, however, uses synthesized baselines 166 based upon
machine configuration data 160 input by a user and a baseline
specification database 158 derived using empirical and analytical
methodologies. In some embodiments, the automated diagnostic system
may process the actual readings 310 and determine that the
statistical average baselines 314 are not available. In such case,
the system 302 may instead request synthesized baselines 166 be
generated by the baseline synthesizer 152.
Machine Fault Screening Method
FIG. 7 shows a flow chart of a process 400 for screening a machine
for vibration faults. At step 402, the machine train, components,
machine speeds, and test locations 304 are identified for a given
test 306 on a given machine. Such test locations 304 and frequency
ranges may be input to the baseline synthesizer 152 and an
automated diagnostic screening system 302. At step 404, the
diagnostic system is configured for screening based upon the
component codes 154 and faults codes 156 obtained from the operator
inputs 160. For example, frequency ranges, sampling rates and other
parameters 308 may be identified. At step 406, vibration readings
310 are obtained by the automated diagnostic system 302. At step
408, the baseline synthesizer 152 executes the method 200 to
synthesize baselines for the machine test. At step 410 the actual
readings 310 are received, transformed and processed to derive
frequency data that may be compared with the synthesized baselines
166. The derived frequency data and the synthesized baselines are
analyzed to perform a desired screening operation, such as to
screen for a select set of fault conditions. At step 412 the
analysis results are output.
One of ordinary skill will appreciate that the order of these steps
may vary. For example, all of the steps pertaining to obtaining the
synthesized baselines may be performed prior to any testing and
stored in a database for later retrieval by the automated
diagnostic system. In another embodiment, the steps pertaining to
obtaining the synthesized baselines may be performed before or
concurrent with obtaining the actual data measurements. For
example, the automated diagnostic system may send a request to the
baseline synthesizer at any step during the process of testing a
machine. Accordingly, in some embodiments the location, frequency
range, component codes and fault codes may be indicated by the
automated diagnostic system, or retrieved from memory.
It is to be understood that the foregoing illustrative embodiments
have been provided merely for the purpose of explanation and are in
no way to be construed as limiting of the invention. Words used
herein are words of description and illustration, rather than words
of limitation. In addition, the advantages and objectives described
herein may not be realized by each and every embodiment practicing
the present invention. Further, although the invention has been
described herein with reference to particular structure, materials
and/or embodiments, the invention is not intended to be limited to
the particulars disclosed herein. Rather, the invention extends to
all functionally equivalent structures, methods and uses, such as
are within the scope of the appended claims. Those skilled in the
art, having the benefit of the teachings of this specification, may
affect numerous modifications thereto and changes may be made
without departing from the scope and spirit of the invention.
* * * * *